The chemokine CCL5 regulates the in vivo cell surface expression of its receptor, CCR5

The efficiency of CCR5 as an HIV coreceptor is strongly dependent on its level of cell surface expression. Therefore, it is of major importance to identify the factors that regulate cell surface density of CCR5. Among the chemokines that bind to CCR5, and induce its internalization, CCL5 is the most abundant in vivo. We show that the level of CCL5 production is a main factor determining CD4+ T cell surface CCR5 density.

We have previously observed that the mean number of CCR5 molecules at the surface of peripheral blood CD4+ T cells is constant over time for a given individual, but varies among individuals ^[1]. We have also shown that this cell surface CCR5 density strongly determines the efficiency of CCR5 as an HIV coreceptor, and that cells expressing a high number of surface CCR5 molecules produce 80-fold more virions after a single round of infection than cells expressing six-fold less CCR5 molecules ^[2]. Consequently, we have reported that, among HIV-1-infected subjects, high CCR5 expressors present with high virus loads ^[1,3], progress faster ^[4,5], and respond less to antiretroviral treatment ^[4]. Moreover, the efficiency of CCR5 as a chemokine receptor is also determined by its level of cell surface expression ^[6]. We investigated the factors that determine the level of cell surface CCR5 expression. Among the chemokines that bind to CCR5 and induce its internalization, CCL5 is the most abundant in human plasma. Consequently, we hypothesized that the amount of CCL5 produced by peripheral blood mononuclear cells (PBMC) could influence CD4+ T cell surface CCR5 density by determining the proportion of CCR5 molecules that are internalized. Indeed, in 18 healthy volunteers, we observed an inverse correlation between the level of CCL5 mRNA expression in their PBMC and the mean number of CCR5 molecules at the surface of their CD4+ T cells (r = 0.502, P = 0.024, data not shown).

To check that the level of CCL5 production actually modulates CD4+ T cell surface CCR5 density, we analyzed in vitro the effect of an increase in the amount of CCL5 produced by PBMC on CD4+ T cell surface CCR5 density. For this purpose, we transduced PBMC with a lentiviral vector harbouring CCL5, or EGFP as a control. CCL5 concentration in cell supernatants 3 days after gene transfer was 3.0 ng/ml and 0.3 ng/ml repectively. Quantitative flow cytometry showed that CCR5 was downregulated at the surface of CCL5-transduced PBMC compared to EGFP-transduced PBMC (Fig. 1a). Thus, in vitro, the overproduction of CCL5 by PBMC resulted in a decrease in CD4+ T cell surface CCR5 density.

Fig. 1

To further test our hypothesis, we monitored CD4+ T cell surface CCR5 density in healthy volunteers who received the CCR5 antagonist vicriviroc (formerly SCH-D). Vicriviroc does not induce CCR5 internalization per se, but inhibits CCR5-CCL5 interaction. We reasoned that the antagonist should interrupt CCL5-induced CCR5 internalization, and thereby cause an increase in cell surface CCR5 expression. We studied 24 volunteers who received 25 mg or 50 mg of vicriviroc (n = 12) or a placebo (n = 12). Twenty-four hours after administration of vicriviroc, we noted an increase in CCL5 plasma concentration [arithmetic means of 22 ng/ml (95% confidence interval [CI] = 12-32) and 39 ng/ml (95% CI = 18, 59), at 0 h and 24 h, respectively, P = 0.006, data not shown]. By contrast, CCL5 plasma level was unchanged in placebo-treated volunteers (data not shown). This increase was not due to an overproduction of CCL5 because CCL5 mRNA expression in PBMC was constant over this 24-h period (data not shown). Vicriviroc actually induced an increase in CD4+ T cell surface CCR5 density in vivo [arithmetic means of 5386 molecules/cell (95% CI = 3860-6911) and 7466 molecules/cell (95% CI = 5676-9257), at 0 h and 24 h, respectively, P = 0.002, Fig. 1b], whereas CD4+ T cell surface CCR5 density was stable in the volunteers who received the placebo [arithmetic means of 7934 molecules/cell (95% CI = 5849-10 019) and 8101 molecules/cell (95% CI = 6146-10 056), at 0 h and 24 h, respectively, P = 0.733, data not shown]. Vicriviroc also increased CCR5 density at the surface of CD8+ T cells [arithmetics means of 10 271 molecules/cell (95% CI = 5998-14 644) and 14 058 molecules/cell [95% CI = 9301-18 815), at 0 h and 24 h, respectively, P = 0.001, data not shown]. Moreover, the effect of vicriviroc on CCR5 expression was specific because it had no effect on CD4+ T cell surface CXCR4 density [arithmetic means of 1323 molecules/cell (95% CI = 1110-1536) and 1576 molecules/cell (95% CI = 1205-1948), at 0 h and 24 h, respectively, P = 0.193, data not shown]. The increase of CCR5 density was not due to an effect of viriviroc on CCR5 transcription because CCR5 mRNA measured by real-time polymerase chain reaction remained stable (data not shown), but rather to a redistribution of the CCR5 receptors between the inner and the outer of the cell because the total (membrane and intracellular) amount of CCR5 molecules present in and on permeabilized CD4+ T cells remained unchanged [arithmetic means of 10 638 molecules/cell (95% CI = 8754-12522) and 10 655 molecules/cell (95% CI = 8115-13195), at 0 h and 24 h, respectively, P > 0.999, Fig. 1c].

Our results are consistent with previous studies reporting that, on exposure to CCL5, 70% of cell membrane CCR5 receptors undergo endocytosis ^[7,8]. It is tempting to try to manipulate CCL5 expression to regulate cell surface CCR5 expression, negatively in case of HIV-1 infection, and positively in case of a disease where CCR5 plays a positive role. Thus, molecules reported to induce CCL5 expression, such as the μ-opioid agonist DAMGO ^[9], thrombin ^[10] or dioxin ^[11], or to inhibit it, such as glatiramer acetate ^[12], progestin ^[13], fibrates ^[14] or indirubin ^[15], might be of interest. Our finding that CCL5 regulates the cell surface expression of its receptor might apply to other chemokines and to other G protein-coupled receptors.

Acknowledgements

We are grateful to the persons who volunteered for this study. This study was supported by SIDACTION.